LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loop

TL;DR

This paper introduces LSUN, a large-scale image dataset created through a semi-automated, human-in-the-loop labeling process, significantly enhancing training data for visual recognition models.

Contribution

The paper presents a novel semi-automated labeling scheme combining deep learning and human effort to efficiently construct a large-scale, diverse image dataset called LSUN.

Findings

Training convolutional networks on LSUN improves accuracy

The dataset contains around one million images per category

The labeling process is effective and scalable

Abstract

While there has been remarkable progress in the performance of visual recognition algorithms, the state-of-the-art models tend to be exceptionally data-hungry. Large labeled training datasets, expensive and tedious to produce, are required to optimize millions of parameters in deep network models. Lagging behind the growth in model capacity, the available datasets are quickly becoming outdated in terms of size and density. To circumvent this bottleneck, we propose to amplify human effort through a partially automated labeling scheme, leveraging deep learning with humans in the loop. Starting from a large set of candidate images for each category, we iteratively sample a subset, ask people to label them, classify the others with a trained model, split the set into positives, negatives, and unlabeled based on the classification confidence, and then iterate with the unlabeled set. To assess the effectiveness of this cascading procedure and enable further progress in visual recognition research, we construct a new image dataset, LSUN. It contains around one million labeled images for each of 10 scene categories and 20 object categories. We experiment with training popular convolutional networks and find that they achieve substantial performance gains when trained on this dataset.